We identified two themes and seven subthemes from the discussions (Figure).
Theme 1: Prediction of the imminence of death
According to the participating physicians, prediction of the imminence of death is important, its main purpose being to enable timely provision of tailored palliative care to patients. By identifying patients who have a limited life expectancy, physicians and care teams can facilitate the proactive evaluation of medical treatment and care in relation to patients’ preferences for the last phase of life.
Well, yes, predicting imminent death forces you as a treatment team to not only look at the next available treatment line, but to critically look at its benefits for a patient who may not have long to live. It is important for physicians to do so because patients might only be focused on that next treatment line. ( Participant 5, medical oncologist )
When predicting imminent death in patients with cancer or COPD, participants rely mainly on their clinical experience and on their knowledge regarding the patient’s disease stage or tumor type. In an open-ended question in the online survey, participants mentioned several clinical factors or symptoms they use to identify a patient’s last phase of life. Those factors are for example performance status for patients with cancer, and acute exacerbations for patients with COPD (Table). Some participants found it difficult to predict imminent death for patients with tumor types for which multiple lines of systemic therapy are available (e.g. breast cancer), because imminence of death may only become evident in case of an acute deterioration after exhausting those treatment lines. Participants who attended patients with cancer as well as patients with COPD found it more difficult to predict imminent death in patients with COPD than in patients with cancer, although for some types of cancer it can be particularly difficult as well.
I think the years of experience have made me better in prognostication, even in situations where the trajectory is different from what we expected. It is never easy and never 100% though. ( Participant 15, medical oncologist )
With breast cancer, it remains difficult because people can live another 10 years with only a few skeletal metastases. ( Participant 13, medical oncologist )
Most participants thought the SQ is a useful tool to support the identification of patients’ last phase of life. They are typically triggered to use the SQ when they notice significant deterioration of a patient’s condition. Facilitators for the use of the SQ are that it is a simple question, clearly formulated, and directly raises awareness about a patient’s imminent death. Additionally, the SQ is recognizable and, therefore, applicable for patients with various chronic diseases. Physicians base their response to the SQ on a combination of intuition, and patient and disease characteristics.
You do have a certain idea of a patient and you wonder, “I am curious if he will make it”. It is a gut feeling, whether you will say yes or no. ( Participant 12, general practitioner )
Well, we have been discussing the surprise question extensively at the department this afternoon. I think that the surprise question itself, however subjective it may be, is not so bad in all its simplicity. ( Participant 8, anesthesiologist )
Participants were not acquainted with other tools than the surprise question to predict death. Participants’ opinions were divided on whether it is preferable to use one’s own subjective clinical judgement or an objective prognostic tool that combines clinical factors. They thought a prognostic tool may give more accurate predictions than subjective judgement, but a tool would probably also require extra time and effort to complete. All participants disagreed with a statement that use of a prognostic tool would make care impersonal and distant.
If you would have tools to estimate it [imminent death] more reliably, that would help. However, I am very curious if that is possible. ( Participant 15, medical oncologist )
Theme 2: Disclosure of the imminence of death
All participants thought that acknowledgment of a patient’s imminent death is important for the initiation of a discussion with patients about their preferences and needs for medical care in the last phase of life. Most participants thought it is useful and feasible to start this discussion early, that is, about one year before a patient’s expected death. This timing may then provide patients and relatives with sufficient time to prepare for the last stage of life and make all necessary arrangements. Some participants, however, thought that one year might be too early to initiate those discussions; they preferred to open such a discussion in the period ‘during which palliative care is actually required’, the period ‘during which maintaining quality of life outweighs prolonging life’, or on the last 6 months of life. Physicians should disclose information about a patient’s imminent death gradually, preferably during multiple conversations, because that gives the patient the opportunity to process the information and to think about preferences for care. Some participants link the timing of those discussions to a significant deterioration of the patient’s disease (e.g. progression of metastases or acute COPD exacerbation), or to the discussion of preferences about resuscitation. Other participants thought that disclosing the imminence of death during those moments may increase the patient’s anxiety or panic.
The period of one year has something arbitrary. A trajectory of 1 year is maybe meaningful because the patient can get used to the last phase and make necessary arrangements. ( Participant 1, pulmonologist )
The last phase of life is an artificial border you draw for yourself. That border, whether 6 months or 1 year, has a different value for each patient. ( Participant 3, medical oncologist )
It is much harder for patients with COPD who are in acute situations admitted to the hospital, but feel perfectly fine when they are discharged and at home. It is then more difficult to talk about such serious topics [such as death]. My experience with COPD patients is that there is a lot of fear and panic. ( Participant 11, pulmonologist )
All participants agreed that the treating physician should be responsible for prognostication and disclosure of a patient’s imminent death. Although the GP could also play a role, the medical specialist knows best when active treatment options for the patient are exhausted, and thus has more insight whether or not the patient’s death is imminent. Participating GPs emphasized that the medical specialist should inform the GP early about the exhaustion of treatments and the imminence of death. Thereafter, the GP should gradually take more responsibility in the further exploring and realizing patients’ preferences for end-of-life care.
I appreciate clear prognostication from the medical specialist. For example, if the medical specialist says that there are no more treatments available for the patient, because it is difficult for me to remain knowledgeable of all treatment options. ( Participant 4, general practitioner )
Participants mentioned several barriers for the disclosure of the imminence of death to patients. First, some participants were concerned about a false prediction of imminent death. They found the SQ to be subjective and difficult to answer. Wrong predictions can emotionally harm patients who strictly hang on to those predictions. A few participants mentioned that they had become more reluctant with their predictions due to experiences with patients whose diseases had followed another course than expected
In all those years, I have sent five people home and said, “You will die within a few days”. All five people were still alive after a year. So you can make huge mistakes. ( Participant 2, medical oncologist )
Second, participants feared that full disclosure of the imminence of death might deprive the patient of hope, especially in patients for whom it is important to maintain hope until the end. Additionally, they were concerned that discussion of the imminence of death may trigger fear in patients or let patients think that the physician is giving up on them. Therefore, they believed that their answer to the SQ should not be disclosed to the patient, unless it is clear that the patient appreciates such disclosure and can cope with it.
Of course, there are always several sides to take into account. Look, you are talking about hope. You also regularly see that people have false hope until the very last chemotherapy, because both the doctor and the patient do not want to talk about the patient’s imminent death. ( Participant 8, pain specialist )
I do not want to invoke a lot of fear because of my answer to the surprise question. I do not want to take away hope from patients by telling them they have one year to live, while that could be five years. ( Participant 2, medical oncologist )
Third, some participants found it difficult to accept imminent death of patients with whom they have an established and good physician–patient relationship, or fear that discussions about patients’ imminent death may affect that relationship. On the other hand, a good physician–patient relationship sometimes makes it easier to initiate the discussion about imminent death.
The more you have a relationship with a patient, the more you don't want to see the end coming. That is a major pitfall. ( Participant 15, medical oncologist )
Lastly, participants felt reluctant to use the SQ and to discuss the imminence of death due to concerns about a lack of palliative care services in their clinical practices. Not all hospitals have specialized palliative care teams that can support patients in the last phase of life. Additionally, lack of time during an outpatient consultation makes the initiation of those discussions difficult.
However, the most difficult thing is to start that conversation [about the last phase of life]. I do not think it gets easier. That is also because I do not have a checklist for it and I have [limited] time at the outpatient clinic. ( Participant 5, medical oncologist )